Abstract:MicroRNAs (miRNAs) are a class of non-coding RNAs about similar to 22nt nucleotides. Studies have proven that miRNAs play key roles in many human complex diseases. Therefore, discovering miRNA-disease associations is beneficial to understanding disease mechanisms, developing drugs, and treating complex diseases. It is well known that it is a time-consuming and expensive process to discover the miRNA-disease associations via biological experiments. Alternatively, computational models could provide a low-cost and high-efficiency way for predicting miRNA-disease associations. In this study, we propose a method (called DNRLMF-MDA) to predict miRNA-disease associations based on dynamic neighborhood regularized logistic matrix factorization. DNRLMF-MDA integrates known miRNA-disease associations, functional similarity and Gaussian Interaction Profile (GIP) kernel similarity of miRNAs, and functional similarity and GIP kernel similarity of diseases. Especially, positive observations (known miRNA-disease associations) are assigned higher importance levels than negative observations (unknown miRNA-disease associations).DNRLMF-MDA computes the probability that a miRNA would interact with a disease by a logistic matrix factorization method, where latent vectors of miRNAs and diseases represent the properties of miRNAs and diseases, respectively, and further improve prediction performance via dynamic neighborhood regularized. The 5-fold cross validation is adopted to assess the performance of our DNRLMF-MDA, as well as other competing methods for comparison. The computational experiments show that DNRLMF-MDA outperforms the state-of-art method PBMDA. The AUC values of DNRLMF-MDA on three datasets are 0.9357, 0.9411, and 0.9416, respectively, which are superior to the PBMDA's results of 0.9218, 0.9187, and 0.9262. The average computation times per 5-fold cross validation of DNRLMF-MDA on three datasets are 38, 46, and 50 seconds, which are shorter than the PBMDA's average computation times of 10869, 916, and 8448 seconds, respectively. DNRLMF-MDA also can predict potential diseases for new miRNAs. Furthermore, case studies illustrate that DNRLMF-MDA is an effective method to predict miRNA-disease associations.

Identifying and Exploiting Potential Mirna-Disease Associations with Neighborhood Regularized Logistic Matrix Factorization

DNRLMF-MDA: Predicting Microrna-Disease Associations Based on Similarities of Micrornas and Diseases.

SNMDA: A novel method for predicting microRNA-disease associations based on sparse neighbourhood.

Matrix Reconstruction with Reliable Neighbors for Predicting Potential MiRNA-disease Associations

Prediction of miRNA-disease associations by neural network-based deep matrix factorization

LRMCMDA: Predicting Mirna-Disease Association by Integrating Low-Rank Matrix Completion with Mirna and Disease Similarity Information

GLNMDA: a Novel Method for Mirna-Disease Association Prediction Based on Global Linear Neighborhoods

MLMDA: a Machine Learning Approach to Predict and Validate MicroRNA–disease Associations by Integrating of Heterogenous Information Sources

HNMDA: heterogeneous network-based miRNA–disease association prediction

Identifying Potential Mirnas–disease Associations with Probability Matrix Factorization

MicroRNA-disease Association Prediction by Matrix Tri-Factorization

Hessian Regularized [Formula: see text]-Nonnegative Matrix Factorization and Deep Learning for miRNA-Disease Associations Prediction

MCMDA: Matrix Completion for MiRNA-disease Association Prediction

NPCMF: Nearest Profile-based Collaborative Matrix Factorization Method for Predicting Mirna-Disease Associations

SCMFMDA: Predicting Microrna-Disease Associations Based on Similarity Constrained Matrix Factorization.

TLHNMDA: Triple Layer Heterogeneous Network Based Inference for MiRNA-Disease Association Prediction

Identifying Potential Mirna-Disease Associations Based on an Improved Manifold Learning Framework

LRMDA: Using Logistic Regression and Random Walk with Restart for MiRNA-Disease Association Prediction

CFNCM: Collaborative filtering neighborhood-based model for predicting miRNA-disease associations

Integrating Random Walk and Binary Regression to Identify Novel Mirna-Disease Association

Improved Prediction of miRNA-Disease Associations Based on Matrix Completion with Network Regularization